摘要
采用光流法结合基于小波变换的像素级图像融合算法,研究了一种动态目标分割方法。光流场可以看作是带有灰度的像素点在图像平面运动产生的瞬时速度场,算法先以光流法计算出的动态目标瞬时速度场的水平速度分量和垂直速度分量作为初始信息,再利用基于小波变换的融合算法获得动态目标的初始分割,最后对初始分割结果进行图像去噪和图像增强,并最终获得清楚的分割图。实验证明,该方法能够产生良好的目标分割效果。
This paper study a segmentation algorithms of moving objects, based on the Optical Flow and pixel-level image fusion with wavelet transform. . Optical flow field can be seen as the instantaneous velocity field which is bring by the pixel movement in the image. The horizontal and the vertical component of the optical flow were calculated and were used as the initial information. Then, the initial segmentation was gotten by pixel-level image fusion algorithms with wavelet transform. Through the Image denoise and enhancement on initial segmentation, the clear segmentation image is obtained. The experimental result demonstrates that this method can give better results for target segmentation.
出处
《电脑开发与应用》
2009年第12期17-19,共3页
Computer Development & Applications
关键词
目标分割
光流法
小波变换
target segmentation, optical flow, wavelet transform
